Abstract

The Arctic region has exhibited dramatic changes in recent times. Many of these are intimately tied up with synoptic activity, but little research has been undertaken on how the characteristics of Arctic cyclones have changed. This paper presents a comprehensive analysis of Arctic (here defined as the domain north of 70°N) cyclones diagnosed with the Melbourne University cyclone tracking scheme applied to the 40-yr ECMWF Re-Analysis (ERA-40) and the NCEP–NCAR (NCEP1) and NCEP–Department of Energy (DOE) Atmospheric Model Intercomparison Project (AMIP)-II (NCEP2) reanalysis sets (the last two extending to the end of 2006). A wide variety of cyclone characteristics is presented as befits these complex features.

In winter the highest density of cyclones is found between Norway and Svalbard and to the east to the Barents and Kara Seas, and significant numbers are found in the central Arctic. In summer the greatest frequencies are found in the central Arctic. The total number of cyclones identified in the ERA-40 record exceeds those in the two NCEP compilations. The mean size of cyclones shows similar maxima in the central Arctic in both winter and summer. By contrast, the greatest mean system depth in winter (in excess of 8 hPa) is found to the southeast of Greenland, although average depths exceed 6 hPa over a considerable portion of the basin. In summer the deepest cyclones are found in the central portion of the Arctic.

The analysis shows that the total number of cyclones in winter exceeds that in summer, a result in contrast to earlier studies. This difference comes about primarily due to the greater numbers of “open strong” systems in winter in all reanalyses. Cyclones in this category are associated with very active synoptic situations; it is of importance that they be included in cyclone counts but would not be considered in many cyclone identification schemes. Since 1979 neither the ERA-40 nor the NCEP2 sets show significant trends in any of the cyclone variables considered. However, over the entire record starting in 1958 the NCEP1 reanalysis exhibits a significant increase in summer cyclone frequency (due mainly to the increase in closed strong systems). Both NCEP1 and ERA-40 also reveal significant increases in the number of summer closed strong cyclones, as well as in their mean depth and intensity in that season.

Interannual variations in Arctic cyclone numbers are closely related to the Arctic Oscillation (AO) index in the full reanalyses records. An even stronger relationship is found between the AO and the number of deep cyclones. These relationships have still held in the last decade when the AO has returned to more normal values but the summer and fall sea ice extent has continued to decrease.

1. Introduction

The Arctic region has experienced significant variability and change in many physical and environmental parameters in recent times. Rigor et al. (2000) analyzed Arctic surface air temperature (SAT) observations during 1979–97 and found significant warming trends (see also Jones and Moberg 2003 and Zhang 2005, who found the Arctic system to have been warming faster than the global average since the 1960s). Changes in sea ice coverage have been dramatic (e.g., Parkinson and Cavalieri 2002; Cavalieri et al. 2003; Comiso 2006; Serreze et al. 2007; Stroeve et al. 2005; 2007), and the decline has been especially steep since 2002. The extent in September 2007 was the lowest ever recorded in the Arctic (beating the previous record set just two years earlier). Lindsay and Zhang (2005) have argued that the late 1980s and early 1990s could be considered a tipping point during which the ice–ocean system began to enter a new era of thinning ice and increasing summer open water because of positive feedbacks, and Comiso (2006) suggested that a dramatic decade of change may have arrived.

Positive trends in Arctic precipitation have also been documented (e.g., Kattsov and Walsh 2000; Przybylak 2002; Hanesiak and Wang 2005; Pavelsky and Smith 2006) and significant increases of Eurasian river discharge into the Arctic Ocean since the 1930s have been observed (e.g., McClelland et al. 2006; Pavelsky and Smith 2006). Consistent with these surface trends, numerous investigations have revealed noteworthy changes in the circulation and structure of the Arctic Ocean (e.g., Swift et al. 2005). Recent studies have also pointed to a greater intrusion of Atlantic waters into the Arctic, and this “Atlantification” is reflected in many variables (Polyakov et al. 2005; Walczowski and Piechura 2006; Holland et al. 2006). The first of these studies suggested that the Arctic Ocean is in transition toward a new, warmer state.

a. Low frequency variability in the Arctic

When drawing attention to these various signals and manifestations of change it is important to bear in mind that the apparent signal will depend on the period explored. Kahl et al. (1993) analyzed lower-troposphere temperature profiles over the Arctic Ocean during the period 1950–90 and found few of their trends to be statistically significant. The apparent contradiction with the findings referred to above is reconciled by the fact that the Arctic cooled substantially between 1940 and 1970 and has shown a strong warming trend since then. Proshutinsky et al. (1999) have noted that their two-climate (anticyclonic and cyclonic) regime theory may provide clues as to the observed decadal variability of the Arctic Ocean. They also remark that this perspective may reconcile different conclusions among scientists who have analyzed Arctic data obtained during different climate states.

The last three decades or so of Arctic monitoring have been very unusual in that it has coincided with the satellite era, the development of the International Arctic Buoy Programme (IABP), and a number of intensive field campaigns. Hence the Arctic over this period has been observed and analyzed more closely than ever before and any changes could potentially be given more prominence than they warrant in that we have a less comprehensive picture of variability before that epoch. Hence we must be cautious about interpreting the apparent behavior in this part of the record.

Given the presence of low frequency variations in the Arctic climate system it is important to use quality records that are as long as possible before we are able to see present variability in context. Several studies have analyzed long times series of a wide variety of indirect or proxy reflections of the state of the Arctic climate system. Among these we mention those of Overpeck et al. (1997), Polyakov et al. (2003, 2004), Schmith and Hansen (2003), and Divine and Dick (2006). These reveal a rich spectrum of temporal variability with oscillations with periods of up to 60–80 yr (often superimposed on trends) showing up in many of the variables. A number of conceptual models based on the range of feedbacks that are present in the complex Arctic atmospheric, oceanic, and cryospheric domain have been proposed to account for the spectrum of variabilities (e.g., Mysak and Venegas 1998; Gudkovich and Kovalev 2002; Wang et al. 2005; Overland and Wang 2005).

Notwithstanding the difficulties and caveats that low frequency variability presents to the interpretation of Arctic time series, there are many lines of evidence to indicate that the Arctic region is experiencing events that are outside the envelope of “recent” variability. An example of this was the breakup over the period from 2000 to 2002 of the largest Arctic ice shelf, the 3000-yr-old Ward Hunt ice shelf (Mueller et al. 2003). Another piece of “evidence” was provided by Wang et al. (2007) who examined the Arctic-wide warm periods 1920–50 and 1979–present in the Intergovernmental Panel on Climate Change (IPCC) Fourth Assessment Report (AR4) simulations. They found that few models captured the amplitude, timing, or length of the earlier event and suggested that this result implies that the event was associated with intrinsic climate variability, a conclusion broadly consistent with that reached by Bengtsson et al. (2004). By contrast, all models represented the current warm period.

b. The Arctic Oscillation and the North Atlantic Oscillation

The dominant mode of Northern Hemisphere (NH) atmospheric variability is the Arctic Oscillation (AO) whose highest high-latitude loadings lie to the north of Iceland and through the central Arctic. With this structure it would be expected that climate conditions in the Arctic would be associated with the phase of the oscillation. It is not surprising, then, that the vacillations of this hemispheric mode have been observed to impact on conditions within the Arctic (Thompson and Wallace 2000; Deser 2000; Dickson et al. 2000). Rigor et al. (2002) used 20 years (1979–98) of Arctic buoy data to document the decrease in Arctic mean sea level pressure (MSLP) over that period and suggested that at least part of the thinning of Arctic sea ice up to that time could be attributed to the trend in the AO toward the high-index polarity.

Another important mode of NH variability is the North Atlantic Oscillation (NAO). There is still much debate as to whether the AO and NAO are independent features. Thompson and Wallace (1998) have made the persuasive argument that the NAO is not a regional North Atlantic phenomenon in its own right but rather the local signature of the broader, hemispheric-scale AO. This perspective is supported by the analyses of Rogers and McHugh (2002), Quadrelli and Wallace (2004), and Feldstein and Franzke (2006). In this paper we also adopt this viewpoint, and suggest that any Arctic climate links with the NAO would hold similarly for the AO. Hence, for example, the links between Icelandic low cyclone activity with the NAO identified by Serreze et al. (1997) are very similar to those with the AO.

However, the idea that the 1990s pattern of Arctic change is related in a simple fashion to a strengthening of the AO has been cast in doubt by continued warming and decreases in ice extent in spite of the reversal of the trend in the AO in the mid-1990s (Morison et al. 2006; Maslanik et al. 2007). Cohen and Barlow (2005) showed that, while the oscillations may contribute to hemispheric and regional warming for multiyear periods, the large-scale features of the global warming trend over the last 30 years are unrelated to the AO and NAO. Simulations by Teng et al. (2006) showed that, similar to the observations, the AO is the dominant factor explaining the variability of the atmosphere and sea ice over the 1870–1999 period but that the twenty-first century Arctic warming has mainly resulted from the radiative forcing of greenhouse gases and a “polar amplification” type of spatial pattern. They suggested that the AO plays a secondary role in its influence on both surface temperature and sea ice concentration. The picture is made more complex by the fact that many high-latitude climate parameters bear a nonlinear relationship to the AO (Wu et al. 2007). These developments raise the question whether the nature of changes occurring in the Arctic means that the associations identified over past epochs apply now to a lesser extent and whether we are, indeed, entering a new phase for the Arctic system.

c. The International Polar Year and Arctic cyclone behavior

In this work we analyze synoptic systems in the Arctic Basin and address the question of their involvement in the changes being observed. One of the motivations for our study was the establishment of the International Polar Year (IPY). This large scientific program is focused on the Arctic and the Antarctic from March 2007 to March 2009 (Allison et al. 2007). Its timing coincides superbly with the urgent need for polar research, and specifically that associated with the shrinking of Arctic snow and ice and the rapid change in many other parameters as discussed above. Particular issues that will be foci in IPY are associated with polar amplification and “trigger points.” There are a number of physical mechanisms that might dictate that warming in the Arctic would be faster than elsewhere. These include positive ice albedo feedback and the fact that in the Arctic a greater fraction of increased energy goes directly to warming the atmosphere (whereas in the tropics a greater part goes to increasing evaporation). Another relevant factor is that, because the vertical stability is greater than in the midlatitudes, any extra heat is deposited in a more shallow layer and potentially results in greater SAT increase (ACIA 2004). Holland and Bitz (2003) suggest that the polar amplification factor lies between 1.5 and 4.5. (Winton 2006 concluded that the surface albedo feedback is a factor in promoting Arctic amplification, but it is not the dominant one.) The Solomon et al. (2007) assessment, based on the most recent and comprehensive analyses, concludes that the average Arctic temperature has increased at almost twice the global average rate in the past 100 years.

The aim of this study is to present insights into the many aspects of cyclones in the Arctic Basin, and we see this paper as a contribution to IPY. Cyclones play a central role in the complex exchanges between the surface and the atmosphere, and are intimately involved with the large-scale circulation, surface temperature, precipitation, sea ice, and the ocean. High-latitude storm activity has been linked with hemispheric-scale oscillations. Cyclones are also known to play an important part in driving interannual and interdecadal annular mode variations (e.g., Luo et al. 2007a, b; Rivière and Orlanski 2007), and even single storms have been shown to have significant impact on these variations. Furthermore, changes in sea ice dramatically modify the fluxes of sensible and latent heat from the surface to the atmosphere and in turn impact cyclone development (see, e.g., Murray and Simmonds 1995). At the same time, a cyclone is capable of exerting very large stresses on the ice surface and, thus, dramatically change its distribution [Maslanik et al. (1996) linked the decrease in Arctic summer ice cover over 1978–95 to increases in the frequency of low pressure systems over the central Arctic]. Storms in the Arctic Basin are increasingly being recognized as strong influences on the budgets and structure in the upper Arctic Ocean (e.g., Yang et al. 2001, 2004).

All these associations suggest that a new and comprehensive examination of cyclones in the Arctic region is very timely in allowing us a broad view of the climate changes we are seeing in the complex Arctic system. In this paper we present a new climatology of Arctic cyclonic systems using the most recent reanalysis datasets. We also diagnose many aspects of cyclone characteristics (including a stratification by type), as we feel these complex phenomena cannot be adequately described in simple terms. These issues are of particular importance given that many studies have highlighted trends in recent times for midlatitude systems to become fewer in number but on average more intense. We also explore the extent to which Arctic cyclone activity has exhibited trends and the extent to which these can be related to changes in the AO.

2. Cyclonic systems in the Arctic

A comprehensive picture of Arctic climate has recently been presented by Przybylak (2003), and he points out that our understanding of the region has changed dramatically over the last century. In the early part of the twentieth century the “glacial anticyclone theory” was generally accepted. This view changed progressively as observational programs and early synoptic maps showed that the Arctic Basin was host to a range of isobaric systems and intense cyclones year round. We now have the benefit of a wide range of sophisticated data acquisition systems and analysis methods with which new perspectives can be obtained of this remote and hostile domain. One of these systems that has provided an excellent in situ view of surface conditions in the Arctic has been the buoy data available under the IABP (http://iabp.apl.washington.edu/), and these have been valuable in demonstrating the frequent occurrence of extreme storms. An analysis of about half a million 6-hourly buoy positions within the Arctic (defined in this paper as all regions north of 70°N) for the period 1979–2004 reveals that in 1% of the sample the pressure is less than 980.4 hPa and that 1% of the 6-h negative pressure changes exceed 10.3 hPa. As an illustration, Fig. 1 shows the pressure trace of (type ICEX-AIR) buoy 1222 (WMO ID 48601) for the period 4–8 January 2003. In the 24-h period from 0600 UTC 5 January this buoy, located in the vicinity of 81°N, 138°W, experienced a pressure reduction of 25.3 hPa, to 976 hPa. While Arctic Basin cyclones may not attract the same interest as systems in the “storm track” regions farther south, it is clear that the Arctic is, indeed, host to a wide variety of cyclone types, some of which are very intense and greatly influence the sea ice and ocean.

Fig. 1.

Pressure time series at IABP buoy 1222 (type ICEX-AIR, WMO ID 48601) for the period 4–8 Jan 2003.

Fig. 1.

Pressure time series at IABP buoy 1222 (type ICEX-AIR, WMO ID 48601) for the period 4–8 Jan 2003.

Changes in Arctic surface conditions represent important modifications to the forcing of the atmosphere and cyclones in particular. Many investigations have revealed that the atmospheric responses to surface forcings in the tropics and in the midlatitudes are toward the “convective limit” and “advective limit,” respectively (Egger 1977; Webster 1981). Glowienka-Hense and Hense (1992) argue that the stationary response to surface temperature forcing in the Arctic is strongly influenced by a third regime, the “eddy flux limit.” Their results emphasized the importance of cyclonic systems in determining the balances. The distribution of Arctic cyclones is known to be determined by a number of factors, including landmass distribution, topography, sea ice concentration, sea surface temperature gradients, and the location and orientation of baroclinic zones. Given the relative importance of these factors it is not surprising that the characteristics of Arctic cyclones are rather different from their midlatitude counterparts [although, in particular, there is much evidence to show that baroclinic instability is also of importance in the Arctic Basin (e.g., Ledrew et al. 1991)]. One of the earliest analyses of (Eurasian sector) Arctic cyclones was undertaken by Wiese (1924). His estimation of the frequency of fall cyclones (his Figs. 11 and 12) showed very large numbers, particularly in years of light summer sea ice. The first comprehensive investigations of Arctic cyclone activity were undertaken in the 1950s (Keegan 1958; Reed and Kunkel 1960). These early studies were conducted with sparse observational data. Since then a number of significant papers have examined Arctic cyclone activity and explored its relationships with NH climate variability. Taken chronologically, these reflect an increase in our understanding of Arctic and subarctic synoptic activity and its quantification. Among some of the key literature we mention the works of Serreze et al. (1993, 1997), Serreze (1995), Overland et al. (1999), Key and Chan (1999), Brümmer et al. (2000), Gulev et al. (2001), McCabe et al. (2001), and Zhang et al. (2004). The last two of these studies found that the strength and frequency of cyclone systems entering the Arctic region have increased in recent decades.

3. Datasets and methods

a. NCEP1, NCEP2, and ERA-40 reanalysis sets

The meteorological reanalyses that we use as a base for our investigation are the MSLP analyses from the National Centers for Environment Prediction (NCEP)–National Center for Atmospheric Research (NCAR), hereafter NCEP1, (Kalnay et al. 1996); the NCEP–Department of Energy (DOE), hereafter NCEP2, (Kanamitsu et al. 2002); and the 40-yr European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis (ERA-40) (Uppala et al. 2005). These three datasets are archived every 6 h and available on a global 2.5° × 2.5° latitude–longitude grid. The periods of these sets used here are 1 January 1958–31 December 2006, 1 January 1979–31 December 2006, and 1 September 1957–31 August 2002, respectively. Even though these sets are constructed with frozen assimilation models, the quality of the reanalyses does change with time. Many of these changes have the character of discrete jumps associated with new data sources coming on stream. One such jump occurred in 1979 when the greater utilization of satellite products started, as well as being the First Global Atmospheric Research Program Global Experiment (FGGE) year in which a range of new observations were obtained. We commented above that it is of value to use datasets that are as long as possible, but we also wish to make sure data quality is maintained as much as possible. To avoid making a judgment on these tradeoffs we have decided to perform our cyclone analyses on the NCEP1 and ERA-40 for the entire period of their records, as well as for the period starting in 1979. In a similar vein, it is known that the reanalyses have their relative strengths and weaknesses, and for this reason we present some of the results undertaken with all sets.

A number of investigations have been conducted on the veracity of these reanalyses, with special reference to the Arctic. Bromwich and Wang (2005) and Bromwich et al. (2007) concluded that the more recent part of the reanalyses presents a reliable description of the Arctic troposphere and that they represent a powerful tool for climate studies in the polar regions. The reanalyses have been used in comparisons of diagnosed cyclone activity in the North Atlantic (Hanson et al. 2004; Trigo 2006). A number of these investigations have commented on greater discrepancies in cyclone analyses from the sets in the presatellite years. Dell’Aquila et al. (2005) found that the NCEP1 and ERA-40 reanalyses had different representations of the (500 hPa) baroclinic available energy conversion processes in the presatellite period. Of relevance to us here is that they identified particular differences in the Labrador Sea.

b. Large-scale indices of the Northern Hemisphere circulation

c. Cyclone identification and tracking

The cyclone identification and tracking presented here has been undertaken with the Melbourne University cyclone tracking scheme (Simmonds and Murray 1999; Simmonds and Keay 2000a, b). It has been shown to perform well in a number of comparisons (e.g., Leonard et al. 1999; Pinto et al. 2005; Raible et al. 2008). This sophisticated scheme uses the quasi-Lagrangian, rather than Eulerian, perspective of cyclone behavior, which is seen as more appropriate, especially in the Arctic Basin. One of the strengths of such an approach is that we deal directly with the systems that impact weather. Concisely described, the automatic algorithm scans a MSLP pattern for possible “lows” by comparing the Laplacian of pressure (LP), ∇2p, at each grid point to those at neighboring grid points. (We ignore MSLP cyclones at locations where the surface elevation exceeds 1 km, as the MSLP has little dynamic meaning in those circumstances.) If a potential low is so identified, the position of the associated pressure minimum is then located by iterative approximation to the center of the ellipsoid of best fit to the pressure surface. If a closed center cannot be found or does not lie within a very small distance, the routine then searches for an “open depression” (in the manner described in Murray and Simmonds 1991). The Laplacian of the pressure in the vicinity of the center can be taken as a measure of intensity (as defined by Petterssen 1956), and systems that fail to reach a specified minimum intensity are excluded. Lows identified are tested by a “concavity criterion,” which requires that the average value of the Laplacian exceed 0.2 hPa (° latitude)−2 over a radius of 2°. We tag systems as “weak” or “strong” depending on whether ∇2p in the vicinity of the center of a low assumes values of 0.2–0.7 hPa (° latitude)−2 or greater than 0.7 hPa (° latitude)−2, respectively. (We stratify the cyclones into four categories according to the “closed” and “open” and the “weak” and “strong” classes.) As an example, we show in Fig. 2 these various categories of systems identified by the scheme in the NCEP1 synoptic map for 0000 UTC 7 December 2006. [Note that the plot, as for others in this paper, is not drawn with the stereographic projection usually used but rather with the orthonomic projection (Simmonds 2003) centered on the North Pole, which has the advantage of presenting the greatest resolution at the central point.] Solid circles and stars represent strong and weak closed systems, respectively, while unfilled circles and stars indicate the locations of strong and weak open systems. Examination of the plot indicates the importance of the consideration of open and weak systems in understanding the synoptics of the region.

Fig. 2.

Arctic region synoptic systems identified in the NCEP1 MSLP chart for 0000 UTC 7 Dec 2006 (orthonormic projection; contour interval 2.5 hPa). Solid circles and stars represent strong and weak closed systems, respectively, while unfilled circles and stars indicate the locations of strong and weak open systems.

Fig. 2.

Arctic region synoptic systems identified in the NCEP1 MSLP chart for 0000 UTC 7 Dec 2006 (orthonormic projection; contour interval 2.5 hPa). Solid circles and stars represent strong and weak closed systems, respectively, while unfilled circles and stars indicate the locations of strong and weak open systems.

The cyclones identified with the above techniques were then tracked with an algorithm that made an estimate of the new position of each cyclone, calculated the probability of associations between the predicted and actual new positions by estimating a decreasing function of the separation and central pressure differences between the two positions, and, finally, found the most probable combination of associations in groups where each system belonged. Cyclones that did not form part of a track that lasted for at least 1 day were excluded from the analyses to be presented below.

The quasi-Lagrangian approach allows us not only to identify cyclones but, from the local MSLP distribution, to determine their intensity and whether they are open or closed. A number of authors have emphasized the importance of quantifying a multiplicity of statistics for each individual cyclone (e.g., Paciorek et al. 2002; Lim and Simmonds 2002; Simmonds et al. 2003; Zhang et al. 2004; Wang et al. 2006; Wernli and Schwierz 2006; Rudeva and Gulev 2007). Among the multiple indices that the cyclone software package determines is the radius R and the “depth” D (see, e.g., Simmonds et al. 1999; Lim and Simmonds 2007) of each system it identifies. The radius is taken as the weighted mean distance from the cyclone center to the points at which ∇2p is zero around the “edge” of the cyclone, while D for an individual system can be expressed in terms of the product of the LP and the square of the radius. This parameter can be seen as the difference between the pressure at the edge of a cyclone and that at the center [and as such is very similar to the “pressure deficit” used by synopticians (e.g., Nielsen and Dole 1992)]. This index is arguably the most useful cyclone characteristic to quantify strength, and it can be shown to be related to the “circulation” around a system. We have performed our cyclone analyses for the winter [December–February (DJF)] and summer [June–August (JJA)] seasons. [Below when we refer to the winter of a given year, the year given refers to the calendar year that has the majority of the months (i.e., the year of the January and February).]

4. Key aspects of Arctic mean sea level pressure and cyclone behavior

To place the cyclonic activity in context we display in Figs. 3a and 3b the long-term MSLP for winter and summer, respectively, from the NCEP1 dataset for the period 1958–2006. The winter climatology shows a high pressure area in excess of 1020 hPa to the north of the Chukchi Sea and a gradient across the Arctic Basin culminating in pressures of less than 1005 hPa over the Norwegian Sea. Also shown in Fig. 3a (dashed contours) is the distribution of the temporal standard deviation of the individual winter mean pressures. This pattern displays its smallest values (less than 3 hPa) in the Chukchi Sea and deviations in excess of 6 hPa between Greenland and Norway, reflecting the influence of strong winter variability in the North Atlantic to the south. The summer MSLP pattern is much flatter and exhibits a weak low centered on the pole. The variability in that season exhibits a maximum (in excess of 3 hPa) in this region, with values decreasing to less than 2 hPa at the edge of our domain. Figure 3c shows the trend in winter surface pressure calculated for this dataset. Virtually the entire region is subject to decreases, and many of those in the American and Icelandic sectors are significantly different from zero (stippled) (95% confidence level). The trends in summer (Fig. 3d) are also predominantly negative, but the pattern is rather different, with the region of significant trend centered near the pole and covering a large portion of the Arctic Ocean. These trends look strong but should be seen against the large multidecadal variability in the Arctic Basin. To present another perspective on this we show in Fig. 3e the winter MSLP trends calculated over the subperiod 1980–2006. This shows decreases over the American sector, but now predominantly increases over the Eurasian sector. Very few of the trends over this short, more recent period are statistically significant. Similarly for summer, the trends for the 1979–2006 period are more modest than those over the complete period (contrasting Figs. 3f and 3d), and again very few of the trends achieve significance.

Fig. 3.

Mean (a) winter and (b) summer MSLP from the NCEP1 dataset (contour interval 2.5 hPa). The dashed contours (at intervals of 1 hPa) show the standard deviation of pressure of the individual winter means. Trend in pressure (contour interval 0.2 hPa decade−1) for (c) winter and (d) summer. (e), (f) As for (c) and (d) but the trends are calculated from the 1980–2006 and 1979–2006 periods. Stippling in (c)–(f) indicates regions over which the trends differ significantly from zero at the 95% confidence level.

Fig. 3.

Mean (a) winter and (b) summer MSLP from the NCEP1 dataset (contour interval 2.5 hPa). The dashed contours (at intervals of 1 hPa) show the standard deviation of pressure of the individual winter means. Trend in pressure (contour interval 0.2 hPa decade−1) for (c) winter and (d) summer. (e), (f) As for (c) and (d) but the trends are calculated from the 1980–2006 and 1979–2006 periods. Stippling in (c)–(f) indicates regions over which the trends differ significantly from zero at the 95% confidence level.

As a preliminary we present a compilation of the frequency distributions of various key cyclone properties as derived with the techniques described above. Figure 4 shows the winter (shaded) and summer distributions of central pressure (pc), R, D, and LP as determined from cyclones identified in the NCEP2 reanalysis (1980–2006). The mode of the pc winter distribution occurs at 1000 hPa and the distribution is negatively skewed. A number of systems achieved a central pressure of 960 hPa or less, while lows with a central pressure in excess of 1030 hPa can be discerned. The winter distribution of R exhibits positive skew, and most cyclones have radii of between 4° and 6° latitude. Both the winter Laplacian and D have even greater skew; in the case of D the mode occurs at 4 hPa, but a considerable proportion of the identified cyclones have depths in excess of 20 hPa. Perhaps surprisingly, the summer distributions are quite similar, although a few differences in detail are noteworthy. For example, the central pressure does not exhibit the extremes at either end and, similarly, the tail of the D and Laplacian distributions is not quite as long (although the distributions of cyclone sizes are very similar).

Fig. 4.

Winter (shaded) and summer frequency distributions of pc, R, D, and LP as determined in the NCEP2 reanalysis (periods 1980–2006 and 1979–2006, respectively). Histogram bin intervals are 2 hPa, 1 hPa, 0.25° lat, and 0.1 hPa (° lat)−2.

Fig. 4.

Winter (shaded) and summer frequency distributions of pc, R, D, and LP as determined in the NCEP2 reanalysis (periods 1980–2006 and 1979–2006, respectively). Histogram bin intervals are 2 hPa, 1 hPa, 0.25° lat, and 0.1 hPa (° lat)−2.

a. Geographic distribution of cyclone properties

The density of systems [the mean number per analysis found in a 103 (° latitude)2 normalizing area] in the two seasons for the ERA-40 set is presented in Fig. 5. In winter (Fig. 5a) a broad region of cyclone densities exceeding 5 extends over the ocean from northern Norway to Svalbard and to the east to the Barents and Kara Seas. A very localized region of high density is found over Baffin Bay. In the vicinity of the pole the density is a not-insignificant 3, while the half of the Arctic Ocean in the American sector is host to mean cyclone frequencies less than this amount. In summer (Fig. 5b) the pattern is rather different. A region of frequencies in excess of 3 is found in the central Arctic and while there is still a local maximum off northwest Norway it is much more modest.

Fig. 5.

(a), (b) System density [the mean number of cyclones found in a 103 (° lat)2 area per analysis] in the ERA-40 set in winter and summer, respectively [contour interval 1 × 10−3 (° lat)−2]. (c), (d) As for (a) and (b) but for the density of cyclogenesis [contour interval 0.25 × 10−3 (° lat)−2 (day)−1].

Fig. 5.

(a), (b) System density [the mean number of cyclones found in a 103 (° lat)2 area per analysis] in the ERA-40 set in winter and summer, respectively [contour interval 1 × 10−3 (° lat)−2]. (c), (d) As for (a) and (b) but for the density of cyclogenesis [contour interval 0.25 × 10−3 (° lat)−2 (day)−1].

In interpreting these density distributions it is of importance to appreciate where the areas of cyclogenesis are located. In winter the highest rates of genesis [in units of the number of new cyclones in the 103 (° latitude)2 normalizing area per day] exceed 2 immediately to the south of Svalbard, and there is significant cyclogenesis in the Barents Sea (Fig. 5c). Elevated levels of genesis are seen in the vicinity of Baffin Bay, consistent with the feature observed above. Genesis activity is weaker in the American sector, and falls to less than 0.25 in the region to the north of Bering Strait. In summer (Fig. 5d) a more subdued pattern emerges, and the main difference is that regions of greatest genesis have moved south to lie over the relatively warmer regions of Alaska and northern Norway.

b. Mean characteristics of Arctic cyclones

The winter and summer distribution over the Arctic Basin of the mean cyclone size (or radius) in ERA-40 is presented in Figs. 6a,b. There is only modest seasonality in this parameter and the largest cyclones are found in the central Arctic in both seasons. The systems found in the basin are rather similar in size to those farther south (Simmonds 2000) and exceed 5° latitude over a significant portion of the domain. In winter (Fig. 6c) the key depth parameter exceeds 8 hPa to the southeast of Greenland, these large values being associated with the North Atlantic storm track to the south. The fact that there are many influential systems within the Arctic Ocean is supported by the fact that mean depths exceed 6 hPa over a considerable portion of the basin. In summer (Fig. 6d) the region between Greenland and Norway becomes one of much more modest mean cyclone depths, and in this season the greatest depths are found in the broad region centered on the pole.

Fig. 6.

(a), (b) Mean radius of cyclones in the ERA-40 set in winter and summer, respectively (contour interval 0.25° lat). (c), (d) As for (a) and (b) but for the mean depth of cyclones (contour interval 1 hPa). (e), (f) Net flux of cyclones in ERA-40 set in winter and summer [the length of the reference arrow is 20 × 10−3 cyclones (° lat)−1 (day)−1)]. (At each point the cyclone algorithm determines how many cyclones cross a unit distance in the eastward and northward direction in a given time. The mean of these counts can then be presented as a vector plot of cyclone flux)

Fig. 6.

(a), (b) Mean radius of cyclones in the ERA-40 set in winter and summer, respectively (contour interval 0.25° lat). (c), (d) As for (a) and (b) but for the mean depth of cyclones (contour interval 1 hPa). (e), (f) Net flux of cyclones in ERA-40 set in winter and summer [the length of the reference arrow is 20 × 10−3 cyclones (° lat)−1 (day)−1)]. (At each point the cyclone algorithm determines how many cyclones cross a unit distance in the eastward and northward direction in a given time. The mean of these counts can then be presented as a vector plot of cyclone flux)

To complement these displays we show the distribution of the net flux of cyclone tracks. In winter we observe a strong flux associated with storm activity in the Atlantic farther south (Fig. 6e). The flux vectors point into the Barents and Kara Seas and then assume a more northward direction (toward the central Arctic) in the Asian sector. Another branch of significant cyclone flux is in the Greenland Sea and from the Bering Strait. Broadly speaking the vectors depict a counterclockwise rotation in the Arctic Basin. Figure 6f shows that the overall structure is similar in summer, except the magnitudes in the Greenland and Norwegian Seas are much smaller. On the other hand, there is an enhanced north and northeastward flux of cyclones to the north of the Asian landmass. The structures revealed in these flux diagrams are consistent with the location of the regions of high cyclogenesis (Figs. 5c,d).

c. Regional summaries

As a summary of the above findings we present in Table 1 the long-term average statistics of all cyclones north of 70°N for the reanalysis datasets. Shown is the total number of cyclones per analysis as well as those stratified by the four categories discussed earlier. For all sets while the closed-strong category is the most populous, it accounts for less than half of the cyclones identified. This reinforces the point made earlier that cyclones in the Arctic Basin are made up from a broad spectrum of synoptic types and many influential systems are missed if more traditional methods of cyclone identification are used. It will be noticed from Table 1 that the total number of cyclones identified in the ERA-40 record exceeds those in the two NCEP compilations. This is, for the most part, also true for each of the four subclasses, except for the closed-weak systems [Wang et al. (2006) came to a similar conclusion on this last point]. The mean radius and depth of Arctic cyclones identified in the two NCEP reanalyses tend to be greater than those in the ERA-40. As to seasonality, in all datasets and subperiods the total number of cyclones in winter exceeds that in summer. This is not surprising given the strong influence of the Atlantic on cyclones on the edge of our domain (and hence the veracity of such statements depends to a considerable extent on exactly where the “Arctic” boundaries are drawn). However, in the interior of the Arctic Ocean we have seen from our plots that summer cyclone frequencies are somewhat greater than those in winter. For the ERA-40 set in both subperiods there are about 0.6 more cyclones per analysis in winter (very similar to the difference seen in the NCEP2 set), while in the NCEP1 the seasonality is somewhat larger with approximately one extra cyclone per analysis. From the data in Table 1 we can deduce that in ERA-40 and NCEP2 most of the winter excess is associated with the enhancement of the number of open-strong systems. There is little seasonality in the frequency of closed-strong (which might be the only ones counted in some schemes) and open-weak systems, while the number of closed-weak systems is greater in summer. Both divisions of the NCEP1 record present a somewhat different picture in the sense that the open-weak and (to a lesser extent) closed-strong systems show decreases in frequency between winter and summer.

Table 1.

Winter and summer mean of cyclone statistics calculated for the dataset and period indicated. The entries for the four classes of cyclone and the total are the mean numbers found per analysis. The radius, depth, and Laplacian are given in units of degrees lat, hPa, and hPa (° lat)−2, respectively.

Winter and summer mean of cyclone statistics calculated for the dataset and period indicated. The entries for the four classes of cyclone and the total are the mean numbers found per analysis. The radius, depth, and Laplacian are given in units of degrees lat, hPa, and hPa (° lat)−2, respectively.
Winter and summer mean of cyclone statistics calculated for the dataset and period indicated. The entries for the four classes of cyclone and the total are the mean numbers found per analysis. The radius, depth, and Laplacian are given in units of degrees lat, hPa, and hPa (° lat)−2, respectively.

Clearly the predominant baroclinic mechanisms, most active in winter, that drive and maintain cyclonic systems in the midlatitudes are weaker over the Arctic Ocean. The seasonal decrease in sea ice extent during summer presents the potential for greatly enhanced fluxes of sensible and latent heat from the ice-free ocean to the atmosphere (e.g., Glowienka-Hense and Hense 1992; Gultepe et al. 2003). Also in summer baroclinicity is observed on the Siberian shelf (Reed and Kunkel 1960). We finally comment that, while there is little seasonality in the mean radius of systems (as we might have expected from Figs. 6a,b), the D and LP are between about 20% and 30% larger in winter in all datasets. This is another aspect of the characterization of cyclones that may not be captured purely by counting systems.

5. Variability and trends in Arctic cyclone behavior

As summarized earlier, many aspects of Arctic climate have undergone significant changes over recent decades, and for a variety of reasons one would expect the behavior of cyclones in the region to have undergone changes in concert. To begin exploring this matter we show in Fig. 7 the time series of the mean winter and summer Arctic cyclone properties. Figure 7a shows the ERA-40 wintertime series for the number of cyclones per analysis, radius, and depth. The plots show considerable interannual variability, with a suggestion of an increase in cyclone numbers up until the early 1990s and a decrease to the end of the series (2002). The time series for R exhibits a similar (but reversed) low frequency variation with the extreme occurring perhaps a little earlier. The depth shows a great deal of variability with no obvious low frequency behavior. The summertime series present a somewhat different picture (Fig. 7b), and it is difficult to discern a clear signal in the cyclone numbers or radius. By contrast, the mean depth shows a peak in the early 1990s.

Fig. 7.

Time series for the Arctic (a) winter and (b) summer means of the number of cyclones per analysis, depth, and radius in the ERA-40 dataset; (c), (d) As in (a), (b) but for the NCEP1 data. (The units for depth and radius are hPa and ° lat, respectively)

Fig. 7.

Time series for the Arctic (a) winter and (b) summer means of the number of cyclones per analysis, depth, and radius in the ERA-40 dataset; (c), (d) As in (a), (b) but for the NCEP1 data. (The units for depth and radius are hPa and ° lat, respectively)

We made the point earlier of the challenges in using long series of reanalysis products and of the problems in ascertaining the reality of any trends, and that there is value in comparing different sets. Figures 7c,d show the winter and summertime series drawn from the NCEP1 reanalysis. The time series from these two reanalyses from 1979 are very similar, which is reassuring. Overall, before this date most of the time series have rather similar structure, with the notable exception of winter cyclone counts. Cyclones identified in NCEP1 show quite high counts in the early 1960s, and the overall positive trend referred to above in connection with ERA-40 is not apparent here.

As discussed earlier, there have been dramatic changes in sea ice since 2002, and we can see some interesting structures in the NCEP1 reanalysis cyclones after that time. While we can make only broad comments on changes over such a short period, we remark that the mean winter radius in both 2005 and 2006 is greater than at any other time during the record and that the mean depths in these two years are only less than those in 1989 and 1993. In summer the picture is somewhat different in that these two years exhibit quite small values and, while not being extreme, fall to the lower end of the distribution.

To what extent are the trends highlighted in these times series above the noise level? We have determined the least squares line of best fit to the various data and in Table 2 present the derived trends in units of per century, designating those that differ significantly from zero at the 95% (in italics) and 99% (bold) confidence levels. One notices immediately that since 1979 neither the ERA-40 nor the NCEP2 reanalyses show significant trends in any of the cyclone variables considered. Over this period the NCEP1 reanalysis reveals significant trends to more closed-weak cyclones and smaller LP in summer and to larger systems in winter. Over the entire record starting at 1958 the ERA-40 and NCEP1 reanalyses exhibit significant trends in different quantities. For example, NCEP1 shows a significant increase in the total number of summer cyclones (due mainly to the increase in closed-strong systems), but this is not apparent in ERA-40 [note, however, even though the variability in the ERA-40 set (Fig. 7b) is greater, the (positive) slope of the trend line is greater in ERA-40]. However, both sets diagnose significant increases in the summer mean depth and Laplacian. In addition, all the significant trends are for increases in the strong cyclone categories and for decreases in the weak categories. The natures of these category changes are consistent with those in storm track changes observed in shorter records (Serreze et al. 1993) and in enhanced CO2 modeling simulations (e.g., Lambert and Fyfe 2006).

Table 2.

Trends in winter and summer means of cyclone statistics calculated for the dataset and period indicated. The entries for the four classes of cyclones and the total are the trends in mean numbers found per analysis per century. The trends for radius, depth, and Laplacian are given in units of ° lat per century, hPa per century, and hPa (° lat)−2 per century, respectively. The trends are determined from the least squares line of best fit to the data, and those trends that differ significantly from zero at the 95% (italics) and 99% (bold) confidence level are highlighted.

Trends in winter and summer means of cyclone statistics calculated for the dataset and period indicated. The entries for the four classes of cyclones and the total are the trends in mean numbers found per analysis per century. The trends for radius, depth, and Laplacian are given in units of ° lat per century, hPa per century, and hPa (° lat)−2 per century, respectively. The trends are determined from the least squares line of best fit to the data, and those trends that differ significantly from zero at the 95% (italics) and 99% (bold) confidence level are highlighted.
Trends in winter and summer means of cyclone statistics calculated for the dataset and period indicated. The entries for the four classes of cyclones and the total are the trends in mean numbers found per analysis per century. The trends for radius, depth, and Laplacian are given in units of ° lat per century, hPa per century, and hPa (° lat)−2 per century, respectively. The trends are determined from the least squares line of best fit to the data, and those trends that differ significantly from zero at the 95% (italics) and 99% (bold) confidence level are highlighted.

a. Associations with the AO and NAO

The trends discussed above reveal a complex picture. We have pointed out that there is a wide range of modes of variability in the Arctic (some with very long periods), and it is important to bear that in mind when examining short records. One question that arises from our analysis is to what extent is the absence of significant trends since 1979 associated with the short record (and hence has fewer degrees of freedom) as opposed to modulation by low frequency variability.

As discussed earlier many studies have identified the important role played by the AO or NAO in inducing variability over the high northern latitudes in the last decades of the twentieth century. To what extent do we see similar connections with Arctic Basin cyclones, and to what degree are the apparent changes in their behavior associated with variations in these large-scale oscillation indices? The indices of the AO and NAO assumed significant positive anomalies in the period up to the mid-1990s, and have since taken on values more typical of the last five decades. As an indication of the variations in the AO and of the associations with Arctic cyclones, Fig. 8 shows the time series of the summer AO and the number of cyclonic systems per analysis. The variations in cyclone counts are similar, particularly in the second half of the record (the correlations over the entire period and from 1979 are 0.64 and 0.69, respectively, both significant at the 99% confidence level). We have correlated the time series of the AO and NAO indices with those of the cyclone counts, the mean depth and mean radius. The magnitude of the correlations with the AO were almost always greater (and in most cases appreciably greater) than those with the NAO. Accordingly, below we only consider the correlations with the former index. Table 3 shows that the number of systems per analysis is positively correlated with the AO for all the datasets and periods. All of the five summer correlations are significantly different from zero (99% level). In winter the relationship is not quite as strong, but there are significant associations for the full ERA-40 (99%) and NCEP1 (95%) records. (Note that all the other correlations, indeed all the correlations in Table 3, are positive.)

Fig. 8.

Time series of the mean summer AO (squares) and the number of cyclonic systems per analysis (circles) in the ERA-40 dataset.

Fig. 8.

Time series of the mean summer AO (squares) and the number of cyclonic systems per analysis (circles) in the ERA-40 dataset.

Table 3.

Correlations of the AO with the number of cyclones per analysis for the datasets and periods indicated. D50 and D10 refer to the time series of system depths that exceed the 50th and 10th percentile determined over the entire record. Coefficients expressed in italics and bold are significantly different from zero at the 95% and 99% confidence levels.

Correlations of the AO with the number of cyclones per analysis for the datasets and periods indicated. D50 and D10 refer to the time series of system depths that exceed the 50th and 10th percentile determined over the entire record. Coefficients expressed in italics and bold are significantly different from zero at the 95% and 99% confidence levels.
Correlations of the AO with the number of cyclones per analysis for the datasets and periods indicated. D50 and D10 refer to the time series of system depths that exceed the 50th and 10th percentile determined over the entire record. Coefficients expressed in italics and bold are significantly different from zero at the 95% and 99% confidence levels.

The mean depth of systems exhibits a somewhat different pattern with the correlations for both seasons and both full reanalyses significant at the 99% level, but the associations in the three shorter records do not achieve significance. The mean radius has an even weaker relationship with the AO with only the winter and summer radii determined from the full NCEP1 period exhibiting significant correlations. The relative weakness of the interannual relationship between the AO and the depth is perhaps surprising given the discussion above with respect to that dynamic considerations suggest that the depth is perhaps the single most useful metric of cyclone influence. However, the relationships between the variables in the Arctic climate system are strongly nonlinear, and one might expect stronger associations with the subset of more active systems, of the sort seen in the North Pacific and Atlantic Oceans (Simmonds and Keay 2002). Following the approach of Simmonds and Keay (2002), we have extracted from our winter and summer compilations the deepest 50% of all cyclones in the Arctic Basin in the various datasets. (The threshold to which this corresponds in the full ERA-40 record is 4.9 and 3.9 hPa in winter and summer, respectively. The threshold values in the other sets are very similar.) We then form a time series of the number of cyclones (D50) that exceed this threshold. In a similar fashion we derive time series for the number of cyclones that are in the deepest 10 percentile (D10) (winter and summer thresholds of 11.5 and 8.8 hPa in ERA-40). Table 3 shows that the number of systems that fall in a given year into the overall deepest half of the distribution is very strongly related to the AO. In fact, all the correlations are significant, and at the 99% level in all but one case. eight of the 10 correlations with the various D10 sets are significant, although their magnitude is a little smaller. Overall, the results in Table 3 suggest that the positive phase of the AO is associated with more numerous, deeper, and (to a much more modest extent) larger cyclonic systems in the Arctic region. These associations are consistent with the results of Rogers et al. (2001), which show annual excess precipitation over evaporation in the Arctic Basin (70°–90°N) to be highly correlated with the NAO.

These changes in cyclone numbers and depths obviously have important implications for changes in Arctic atmosphere–surface interactions. The increases in net cyclone influence would have deleterious effects on the integrity of the sea ice in affected sectors by hastening mechanical breakup with, again, implications for changes in the concentration of the sea ice that persists throughout the year. There is much evidence that even single Arctic synoptic systems can greatly influence sea ice distribution (Simmonds and Drinkwater 2007).

6. Comparison of cyclone behavior with earlier studies

In broad terms our results confirm the findings in earlier studies that the most synoptically active region of the Arctic lies on the Eurasian side, particularly in winter. However, comprehensive comparison with other climatologies is difficult in that low frequency variability in the Arctic means that conclusions drawn may be quite specific to the period examined. For example, Przybylak (2003) has pointed out that the cyclone analysis of Serreze and Barry (1988), who analyzed the period 1979–85, showed few lows in the Baffin Bay region, a structure considerably at variance with that of other studies using different periods.

An analysis of Arctic cyclone activity up till 2002 was undertaken by Zhang et al. (2004) using the NCEP1 dataset. They found Arctic (defined as in the present paper) cyclone activity to have increased during the second half of the twentieth century, consistent with the positive trends that we have documented in the number of strong systems. They also found that the number and intensity of cyclones entering the Arctic from the midlatitudes has increased (particularly in summer) and that Arctic cyclone activity displays significant low frequency variability, with low activity in the 1960s and enhanced activity in the 1990s. They identified higher frequency oscillations on top of these low frequencies, one of which (7.8 yr) corresponds to the alternation of the cyclonic and anticyclonic regimes of the Arctic Ocean and sea ice motions.

As discussed earlier, the Arctic region hosts a number of mechanisms that can produce or assist cyclogenesis. The relative importance of these factors changes with season, so a priori it is not obvious how cyclone characteristics might change from winter to summer. Serreze et al. (1993) identified more numerous but weaker Arctic (in their case 65°–90°N) cyclones in summer than in winter. Brümmer et al. (2000) identified Arctic (north of 60°N) cyclones in five years of an earlier ECMWF reanalysis and found about 20% more cyclones in summer than winter. Zhang et al. (2004) commented that their cyclone trajectory count displays modest seasonality but did find that winter Arctic cyclones tended to be more intense, shorter lived, and fewer in number than their summer counterparts.

Our investigation has shown cyclones to be more numerous in summer in the central and American sectors of the Arctic, while the opposite is true over the rest of the basin. The seasonality clearly depends on how cyclones are counted, and we have argued on the importance of considering all categories of cyclones, including whether the features are open or closed and strong or weak. The counts presented in Table 1 imply that closed systems account for only between 44% and 61% of the total number of lows identified (depending on the analysis, period, and season). When we confine our attention to these systems, our compilation with the ERA-40 and NCEP2 sets reveals that about 10% more systems are identified in summer, in broad agreement with the literature results cited above (although there is very little seasonality in the NCEP1 set). In ERA-40 and NCEP2 about 40% more open systems are identified in winter than in summer, and this rises to about 70% for NCEP1. In accord with synoptic experience, this indicates that low pressure troughs and open depressions are a significant part of the Arctic winter synoptic environment. Another interesting statistic that can be deduced from Table 1 is that there are 40% more strong systems identified for winter than summer. Conversely, there are about 15% more weak systems in summer in the ERA-40 and NCEP1 reanalyses. When all systems over our domain are considered, we find winter systems are deeper and more numerous in all our datasets. Table 1 reveals that much of the difference in seasonality can be attributed to the numbers of open-strong systems being more numerous in winter. These features are frequently associated with significant vorticity and are important components of the Arctic “weather” mix and in winter can be associated with troughs extending from outside the immediate Arctic Basin. They may not be included in compilations performed with traditional cyclone counting techniques. [In addition, studies based on older analyses may have had an observational bias militating to the identification of lows in summer. It is conceivable that more lows were found in summer because more MSLP observations (e.g., more ship data) were taken during that season.]

7. Concluding remarks

We have presented a multivariate climatology of Arctic cyclones based on the Melbourne University cyclone tracking scheme and the ERA-40, NCEP1, and NCEP2 reanalysis sets. In winter the highest density of cyclones is found between Norway and Svalbard and to the east to the Barents and Kara Seas, and significant numbers are found in the central Arctic. Cyclones are less numerous in the American sector of the Arctic Ocean. In summer the greatest frequencies are found in the central Arctic. The regions of cyclogenesis also show significant seasonality. In common with many studies we find that the total number of cyclones identified in the ERA-40 record exceeds that in the NCEP1 compilation. The mean size of cyclones show similar maxima in the central Arctic in both winter and summer. By contrast, the greatest mean system depth in winter (in excess of 8 hPa) is found to the southeast of Greenland, although average depths exceed 6 hPa over a considerable portion of the basin. In summer the deepest cyclones are found in the central portion of the Arctic.

In contrast to earlier studies, we find the total number of cyclones in winter to exceed that in summer. This comes about primarily due to the greater numbers of “open-strong” systems in winter in all the reanalyses. There is little seasonality in the frequency of closed-strong systems (which might be the only ones counted in many schemes) and open-weak systems, while the number of closed-weak systems are greater in summer. The open-strong systems are known to be associated with very active synoptic situations, and it is important that they be included in cyclone counts.

Our results reveal high levels of interannual variability of cyclone properties. Since 1979 the variability is similar in the three reanalyses. We note that since 1979 neither the ERA-40 nor the NCEP2 sets show significant trends in any of the cyclone variables considered (while the NCEP1 reanalysis reveals significant trends in some parameters). Over the entire record starting at 1958 the NCEP1 reanalysis exhibits a significant increase in summer cyclone frequency (due mainly to the increase in closed-strong systems), although this is not apparent in ERA-40. Both reanalyses also reveal significant increases in the number of summer closed-strong cyclones, as well as for mean depth and Laplacian.

Correlation analysis indicates that the interannual variations in Arctic cyclone properties are closely associated with the AO and NAO, with the former showing stronger associations. The correlations with the AO are particularly strong in summer. There is only a weak link between mean seasonal cyclone radius and the AO, but a much closer association with depth when the entire period is considered. This is particularly marked when the time series of the deepest half of the overall cyclone pool is considered. These findings reinforce the view that depth represents a very valuable and sensitive measure of mean cyclone behavior. The correlations mentioned above are calculated over a period in which the AO assumed positive anomalies (in the period up to the mid-1990s) and have since assumed values more typical of the last five decades. This relationship has still held in the last decade when the AO has returned to more normal values, but while the summer and fall sea ice extent has continued to decrease.

The links between the AO/NAO and high-latitude climate are complex and highly nonlinear, in some cases, and continue to attract much attention (Hilmer and Jung 2000; Stephenson et al. 2006; Miller et al. 2006). Graversen (2006) has documented increases in atmospheric northward energy transfer across 60°N. He pointed out that this trend was opposite to that of the AO over the last decade, and opined this could be a consequence of changes in the baroclinic structure of the synoptic-scale waves in the Pacific and Atlantic sectors and changes that are not primarily associated with the strength of the zonal flow. The amplification and vertical structure of synoptic-scale waves also depends on the static stability of the atmosphere and the surface friction. The critical horizontal wavelength for baroclinic energy to be released is proportional to the static stability (e.g., Walland and Simmonds 1999; Holton 2004). Hence, even though the AO index may not change, the baroclinic structure of the waves might be altered in such a way that the northward heat transport increases. It has been found that, even though part of variations in the characteristics of the North Atlantic storm track might be associated with the AO, a significant part cannot be linearly attributed to this mode (e.g., Chang and Fu 2003). Graversen (2006) showed that though part of the SAT trend can be related to the AO in localized parts of the Arctic area, the mean Arctic SAT trend shows no significant linkage to the AO. Many authors have warned against making too-strong interpretations from short records of time series with white or red characteristics (e.g., Wunsch 1999). Recent analyses suggest that the NAO data are not characterized by a random walk process and that its structure points to a complex dynamics, which has yet to be comprehensively explained (e.g., Stephenson et al. 2000; Mills 2004).

As a final comment, we have stressed in this paper the importance of conducting investigations such as the present one with a number of independent datasets, and this presents a very useful way of assessing the robustness of the results. A second uncertainty that we have not addressed here is associated with the choice of the cyclone identification and tracking scheme. Raible et al. (2008) have shown that the sensitivity of their results to the choice of algorithm was significant and clearly points to the desirability of conducting these investigations with detection and tracking schemes that capture the full complexity of these features.

Acknowledgments

Parts of this research were made possible by a grant from the Australian Research Council. The authors express their appreciation to staff at ECMWF and NCEP for making freely available the reanalysis datasets. We also wish to thank Ignatius Rigor for his assistance in obtaining the IABP data. We are grateful for the very constructive comments of two reviewers.

REFERENCES

REFERENCES
ACIA
,
2004
:
Impacts of a Warming Arctic: Arctic Climate Impact Assessment.
Cambridge University Press, 139 pp
.
Allison
,
I.
, and
Coauthors
,
2007
:
The scope of science for the International Polar Year 2007–2008. WMO/TD 1364, World Meteorological Organization, 79 pp
.
Bengtsson
,
L.
,
V. A.
Semenov
, and
O. M.
Johannessen
,
2004
:
The early twentieth-century warming in the Arctic: A possible mechanism.
J. Climate
,
17
,
4045
4057
.
Bromwich
,
D. H.
, and
S-H.
Wang
,
2005
:
Evaluation of the NCEP–NCAR and ECMWF 15- and 40-yr reanalyses using rawinsonde data from two independent Arctic field experiments.
Mon. Wea. Rev.
,
133
,
3562
3578
.
Bromwich
,
D. H.
,
R. L.
Fogt
,
K. I.
Hodges
, and
J. E.
Walsh
,
2007
:
A tropospheric assessment of the ERA-40, NCEP, and JRA-25 global reanalyses in the polar regions.
J. Geophys. Res.
,
112
.
D10111, doi:10.1029/2006JD007859
.
Brümmer
,
B.
,
S.
Thiemann
, and
A.
Kirchgäbner
,
2000
:
A cyclone statistics for the Arctic based on European Centre re-analysis data.
Meteor. Atmos. Phys.
,
75
,
233
250
.
Cavalieri
,
D. J.
,
C. L.
Parkinson
, and
K. Y.
Vinnikov
,
2003
:
30-year satellite record reveals contrasting Arctic and Antarctic decadal sea ice variability.
Geophys. Res. Lett.
,
30
.
1970, doi:10.1029/2003GL018031
.
Chang
,
E. K. M.
, and
Y.
Fu
,
2003
:
Using mean flow change as a proxy to infer interdecadal storm track variability.
J. Climate
,
16
,
2178
2196
.
Cohen
,
J.
, and
M.
Barlow
,
2005
:
The NAO, the AO, and global warming: How closely related?
J. Climate
,
18
,
4498
4513
.
Comiso
,
J. C.
,
2006
:
Abrupt decline in the Arctic winter sea ice cover.
Geophys. Res. Lett.
,
33
.
L18504, doi:10.1029/2006GL027341
.
Dell’Aquila
,
A.
,
V.
Lucarini
,
P. M.
Ruti
, and
S.
Calmanti
,
2005
:
Hayashi spectra of the northern hemisphere mid-latitude atmospheric variability in the NCEP-NCAR and ECMWF reanalyses.
Climate Dyn.
,
25
,
639
652
.
Deser
,
C.
,
2000
:
On the teleconnectivity of the “Arctic Oscillation”.
Geophys. Res. Lett.
,
27
,
779
782
.
Dickson
,
R. R.
, and
Coauthors
,
2000
:
The Arctic Ocean response to the North Atlantic oscillation.
J. Climate
,
13
,
2671
2696
.
Divine
,
D. V.
, and
C.
Dick
,
2006
:
Historical variability of sea ice edge position in the Nordic Seas.
J. Geophys. Res.
,
111
.
C01001, doi:10.1029/2004JC002851
.
Egger
,
J.
,
1977
:
On the linear theory of atmospheric response to sea surface temperature anomalies.
J. Atmos. Sci.
,
34
,
603
614
.
Feldstein
,
S. B.
, and
C.
Franzke
,
2006
:
Are the North Atlantic Oscillation and the Northern Annular Mode distinguishable?
J. Atmos. Sci.
,
63
,
2915
2930
.
Glowienka-Hense
,
R.
, and
A.
Hense
,
1992
:
The effect of an Arctic polynya on the Northern Hemisphere mean circulation and eddy regime: A numerical experiment.
Climate Dyn.
,
7
,
155
163
.
Graversen
,
R. G.
,
2006
:
Do changes in the midlatitude circulation have any impact on the Arctic surface air temperature trend?
J. Climate
,
19
,
5422
5438
.
Gudkovich
,
Z. M.
, and
E. G.
Kovalev
,
2002
:
On the mechanisms of the cyclic climate changes in the Arctic and the Antarctic.
Oceanology
,
42
,
775
780
.
Gulev
,
S. K.
,
O.
Zolina
, and
S.
Grigoriev
,
2001
:
Extratropical cyclone variability in the Northern Hemisphere winter from the NCEP/NCAR reanalysis data.
Climate Dyn.
,
17
,
795
809
.
Gultepe
,
I.
,
G. A.
Isaac
,
A.
Williams
,
D.
Marcotte
, and
K. B.
Strawbridge
,
2003
:
Turbulent heat fluxes over leads and polynyas, and their effects on Arctic clouds during FIRE.ACE: Aircraft observations for April 1998.
Atmos.–Ocean
,
41
,
15
34
.
Hanesiak
,
J. M.
, and
X.
Wang
,
2005
:
Adverse-weather trends in the Canadian Arctic.
J. Climate
,
18
,
3140
3156
.
Hanson
,
C. E.
,
J. P.
Palutikof
, and
T. D.
Davies
,
2004
:
Objective cyclone climatologies of the North Atlantic: A comparison between the ECMWF and NCEP reanalyses.
Climate Dyn.
,
22
,
757
769
.
Hilmer
,
M.
, and
T.
Jung
,
2000
:
Evidence for a recent change in the link between the North Atlantic Oscillation and Arctic sea ice export.
Geophys. Res. Lett.
,
27
,
989
992
.
Holland
,
M. M.
, and
C. M.
Bitz
,
2003
:
Polar amplification of climate change in coupled models.
Climate Dyn.
,
21
,
221
232
.
Holland
,
M. M.
,
J.
Finnis
, and
M. C.
Serreze
,
2006
:
Simulated Arctic Ocean freshwater budgets in the twentieth and twenty-first centuries.
J. Climate
,
19
,
6221
6242
.
Holton
,
J. R.
,
2004
:
An Introduction to Dynamic Meteorology.
4th ed. Academic Press, 535 pp
.
Jones
,
P. D.
, and
A.
Moberg
,
2003
:
Hemispheric and large-scale surface air temperature variations: An extensive revision and an update to 2001.
J. Climate
,
16
,
206
223
.
Kahl
,
J. D.
,
D. J.
Charlevoix
,
N. A.
Zaitseva
,
R. C.
Schnell
, and
M. C.
Serreze
,
1993
:
Absence of evidence for Greenhouse warming over the Arctic Ocean in the past 40 years.
Nature
,
361
,
335
337
.
Kalnay
,
E.
, and
Coauthors
,
1996
:
The NCEP/NCAR 40-Year Reanalysis Project.
Bull. Amer. Meteor. Soc.
,
77
,
437
471
.
Kanamitsu
,
M.
,
W.
Ebisuzaki
,
J.
Woollen
,
S-K.
Yang
,
J. J.
Hnilo
,
M.
Fiorino
, and
G. L.
Potter
,
2002
:
NCEP–DOE AMIP-II Reanalysis (R-2).
Bull. Amer. Meteor. Soc.
,
83
,
1631
1643
.
Kattsov
,
V. M.
, and
J. E.
Walsh
,
2000
:
Twentieth-century trends of Arctic precipitation from observational data and a climate model simulation.
J. Climate
,
13
,
1362
1370
.
Keegan
,
T. J.
,
1958
:
Arctic synoptic activity in winter.
J. Meteor.
,
15
,
513
521
.
Key
,
J. R.
, and
A. C. K.
Chan
,
1999
:
Multidecadal global and regional trends in 1000 mb and 500 mb cyclone frequencies.
Geophys. Res. Lett.
,
26
,
2053
2056
.
Lambert
,
S. J.
, and
J. C.
Fyfe
,
2006
:
Changes in winter cyclone frequencies and strengths simulated in enhanced greenhouse warming experiments: Results from the models participating in the IPCC diagnostic exercise.
Climate Dyn.
,
26
,
713
728
.
Ledrew
,
E. F.
,
D.
Johnson
, and
J. A.
Maslanik
,
1991
:
An examination of atmospheric mechanisms that may be responsible for the annual reversal of the Beaufort sea ice field.
Int. J. Climatol.
,
11
,
841
859
.
Leonard
,
S. R.
,
J.
Turner
, and
A.
Van Der Wal
,
1999
:
An assessment of three automatic depression tracking schemes.
Meteor. Appl.
,
6
,
173
183
.
Lim
,
E-P.
, and
I.
Simmonds
,
2002
:
Explosive cyclone development in the Southern Hemisphere and a comparison with Northern Hemisphere events.
Mon. Wea. Rev.
,
130
,
2188
2209
.
Lim
,
E-P.
, and
I.
Simmonds
,
2007
:
Southern Hemisphere winter extratropical cyclone characteristics and vertical organization observed with the ERA-40 reanalysis data in 1979–2001.
J. Climate
,
20
,
2675
2690
.
Lindsay
,
R. W.
, and
J.
Zhang
,
2005
:
The thinning of Arctic sea ice, 1988–2003: Have we passed a tipping point?
J. Climate
,
18
,
4879
4894
.
Luo
,
D.
,
A. R.
Lupo
, and
H.
Wan
,
2007a
:
Dynamics of eddy-driven low-frequency dipole modes. Part I: A simple model of North Atlantic oscillations.
J. Atmos. Sci.
,
64
,
3
28
.
Luo
,
D.
,
T.
Gong
,
Y.
Diao
, and
W.
Zhou
,
2007b
:
Storm tracks and annular modes.
Geophys. Res. Lett.
,
34
.
L17701, doi:10.1029/2007GL030436
.
Maslanik
,
J. A.
,
M. C.
Serreze
, and
R. G.
Barry
,
1996
:
Recent decreases in Arctic summer ice cover and linkages to atmospheric circulation anomalies.
Geophys. Res. Lett.
,
23
,
1677
1680
.
Maslanik
,
J. A.
,
S.
Drobot
,
C.
Fowler
,
W.
Emery
, and
R.
Barry
,
2007
:
On the Arctic climate paradox and the continuing role of atmospheric circulation in affecting sea ice conditions.
Geophys. Res. Lett.
,
34
.
L03711, doi:10.1029/2006GL028269
.
McCabe
,
G. J.
,
M. P.
Clark
, and
M. C.
Serreze
,
2001
:
Trends in Northern Hemisphere surface cyclone frequency and intensity.
J. Climate
,
14
,
2763
2768
.
McClelland
,
J. W.
,
S. J.
Déry
,
B. J.
Peterson
,
R. M.
Holmes
, and
E. F.
Wood
,
2006
:
A pan-Arctic evaluation of changes in river discharge during the latter half of the 20th century.
Geophys. Res. Lett.
,
33
.
L06715, doi:10.1029/2006GL025753
.
Miller
,
R. L.
,
G. A.
Schmidt
, and
D. T.
Shindell
,
2006
:
Forced annular variations in the 20th century Intergovernmental Panel on Climate Change Fourth Assessment Report models.
J. Geophys. Res.
,
111
.
D18101, doi:10.1029/2005JD006323
.
Mills
,
T. C.
,
2004
:
Is the North Atlantic Oscillation a random walk? A comment with further results.
Int. J. Climatol.
,
24
,
377
383
.
Morison
,
J.
,
M.
Steele
,
T.
Kikuchi
,
K.
Falkner
, and
W.
Smethie
,
2006
:
Relaxation of central Arctic Ocean hydrography to pre-1990s climatology.
Geophys. Res. Lett.
,
33
.
L17604, doi:10.1029/2006GL026826
.
Mueller
,
D. R.
,
W. F.
Vincent
, and
M. O.
Jeffries
,
2003
:
Break-up of the largest Arctic ice shelf and associated loss of an epishelf lake.
Geophys. Res. Lett.
,
30
.
2031, doi:10.1029/2003GL017931
.
Murray
,
R. J.
, and
I.
Simmonds
,
1991
:
A numerical scheme for tracking cyclone centres from digital data. Part I: Development and operation of the scheme.
Aust. Meteor. Mag.
,
39
,
155
166
.
Murray
,
R. J.
, and
I.
Simmonds
,
1995
:
Responses of climate and cyclones to reductions in Arctic winter sea ice.
J. Geophys. Res.
,
100
,
4791
4806
.
Mysak
,
L. A.
, and
S. A.
Venegas
,
1998
:
Decadal climate oscillations in the Arctic: A new feedback loop for atmosphere-ice-ocean interactions.
Geophys. Res. Lett.
,
25
,
3607
3610
.
Nielsen
,
J. W.
, and
R. M.
Dole
,
1992
:
A survey of extratropical cyclone characteristics during GALE.
Mon. Wea. Rev.
,
120
,
1156
1167
.
Overland
,
J. E.
, and
M.
Wang
,
2005
:
The Arctic climate paradox: The recent decrease of the Arctic Oscillation.
Geophys. Res. Lett.
,
32
.
L06701, doi:10.1029/2004GL021752
.
Overland
,
J. E.
,
J. M.
Adams
, and
N. A.
Bond
,
1999
:
Decadal variability of the Aleutian low and its relation to high-latitude circulation.
J. Climate
,
12
,
1542
1548
.
Overpeck
,
J.
, and
Coauthors
,
1997
:
Arctic environmental change of the last four centuries.
Science
,
278
,
1251
1256
.
Paciorek
,
C. J.
,
J. S.
Risbey
,
V.
Ventura
, and
R. D.
Rosen
,
2002
:
Multiple indices of Northern Hemisphere cyclone activity, winters 1949–99.
J. Climate
,
15
,
1573
1590
.
Parkinson
,
C. L.
, and
D. J.
Cavalieri
,
2002
:
A 21 year record of Arctic sea-ice extents and their regional, seasonal and monthly variability and trends.
Ann. Glaciol.
,
34
,
441
446
.
Pavelsky
,
T. M.
, and
L. C.
Smith
,
2006
:
Intercomparison of four global precipitation data sets and their correlation with increased Eurasian river discharge to the Arctic Ocean.
J. Geophys. Res.
,
111
.
D21112, doi:10.1029/2006JD007230
.
Petterssen
,
S.
,
1956
:
Weather Analysis and Forecasting.
Vol. 1, Motion and Motion Systems. 2nd ed. McGraw-Hill, 428 pp
.
Pinto
,
J. G.
,
T.
Spangehl
,
U.
Ulbrich
, and
P.
Speth
,
2005
:
Sensitivities of a cyclone detection and tracking algorithm: Individual tracks and climatology.
Meteor. Z.
,
14
,
823
838
.
Polyakov
,
I. V.
,
R. V.
Bekryaev
,
G. V.
Alekseev
,
U. S.
Bhatt
,
R. L.
Colony
,
M. A.
Johnson
,
A. P.
Maskshtas
, and
D.
Walsh
,
2003
:
Variability and trends of air temperature and pressure in the maritime Arctic, 1875–2000.
J. Climate
,
16
,
2067
2077
.
Polyakov
,
I. V.
, and
Coauthors
,
2004
:
Variability of the intermediate Atlantic water of the Arctic Ocean over the last 100 years.
J. Climate
,
17
,
4485
4497
.
Polyakov
,
I. V.
, and
Coauthors
,
2005
:
One more step toward a warmer Arctic.
Geophys. Res. Lett.
,
32
.
L17605, doi:10.1029/2005GL023740
.
Proshutinsky
,
A. Y.
,
I. V.
Polyakov
, and
M. A.
Johnson
,
1999
:
Climate states and variability of Arctic ice and water dynamics during 1946-1997.
Polar Res.
,
18
,
135
142
.
Przybylak
,
R.
,
2002
:
Variability of total and solid precipitation in the Canadian Arctic from 1950 to 1995.
Int. J. Climatol.
,
22
,
395
420
.
Przybylak
,
R.
,
2003
:
The Climate of the Arctic.
Kluwer Academic, 270 pp
.
Quadrelli
,
R.
, and
J. M.
Wallace
,
2004
:
A simplified linear framework for interpreting patterns of Northern Hemisphere wintertime climate variability.
J. Climate
,
17
,
3728
3744
.
Raible
,
C. C.
,
P. M.
Della-Marta
,
C.
Schwierz
,
H.
Wernli
, and
R.
Blender
,
2008
:
Northern Hemisphere extratropical cyclones: A comparison of detection and tracking methods and different reanalyses.
Mon. Wea. Rev.
,
136
,
880
897
.
Reed
,
R. J.
, and
B. A.
Kunkel
,
1960
:
The Arctic circulation in summer.
J. Meteor.
,
17
,
489
506
.
Rigor
,
I. G.
,
R. L.
Colony
, and
S.
Martin
,
2000
:
Variations in surface air temperature observations in the Arctic, 1979–97.
J. Climate
,
13
,
896
914
.
Rigor
,
I. G.
,
J. M.
Wallace
, and
R. L.
Colony
,
2002
:
Response of sea ice to the Arctic Oscillation.
J. Climate
,
15
,
2648
2663
.
Rivière
,
G.
, and
I.
Orlanski
,
2007
:
Characteristics of the Atlantic storm-track eddy activity and its relation with the North Atlantic Oscillation.
J. Atmos. Sci.
,
64
,
241
266
.
Rogers
,
A. N.
,
D. H.
Bromwich
,
E. N.
Sinclair
, and
R. I.
Cullather
,
2001
:
The atmospheric hydrologic cycle over the Arctic Basin from reanalyses. Part II: Interannual variability.
J. Climate
,
14
,
2414
2429
.
Rogers
,
J. C.
, and
M. J.
McHugh
,
2002
:
On the separability of the North Atlantic Oscillation and Arctic Oscillation.
Climate Dyn.
,
19
,
599
608
.
Rudeva
,
I.
, and
S. K.
Gulev
,
2007
:
Climatology of cyclone size characteristics and their changes during the cyclone life cycle.
Mon. Wea. Rev.
,
135
,
2568
2587
.
Schmith
,
T.
, and
C.
Hansen
,
2003
:
Fram Strait ice export during the nineteenth and twentieth centuries reconstructed from a multiyear sea ice index from southwestern Greenland.
J. Climate
,
16
,
2782
2791
.
Serreze
,
M. C.
,
1995
:
Climatological aspects of cyclone development and decay in the Arctic.
Atmos.–Ocean
,
33
,
1
23
.
Serreze
,
M. C.
, and
R. G.
Barry
,
1988
:
Synoptic activity in the Arctic Basin.
J. Climate
,
1
,
1276
1295
.
Serreze
,
M. C.
,
J. E.
Box
,
R. G.
Barry
, and
J. E.
Walsh
,
1993
:
Characteristics of Arctic synoptic activity, 1952-1989.
Meteor. Atmos. Phys.
,
51
,
147
164
.
Serreze
,
M. C.
,
F.
Carse
,
R. G.
Barry
, and
J. C.
Rogers
,
1997
:
Icelandic low cyclone activity: Climatological features, linkages with the NAO, and relationships with recent changes in the Northern Hemisphere circulation.
J. Climate
,
10
,
453
464
.
Serreze
,
M. C.
,
M. M.
Holland
, and
J.
Stroeve
,
2007
:
Perspectives on the Arctic’s shrinking sea-ice cover.
Science
,
315
,
1533
1536
.
Simmonds
,
I.
,
2000
:
Size changes over the life of sea level cyclones in the NCEP reanalysis.
Mon. Wea. Rev.
,
128
,
4118
4125
.
Simmonds
,
I.
,
2003
:
Regional and large-scale influences on Antarctic Peninsula climate.
Antarctic Peninsula Climate Variability: Historical and Paleoenvironmental Perspectives, E. Domack et al., Eds., Antarctic Research Series, Vol. 79, Amer. Geophys. Union, 31–42
.
Simmonds
,
I.
, and
R. J.
Murray
,
1999
:
Southern extratropical cyclone behavior in ECMWF analyses during the FROST Special Observing Periods.
Wea. Forecasting
,
14
,
878
891
.
Simmonds
,
I.
, and
K.
Keay
,
2000a
:
Mean Southern Hemisphere extratropical cyclone behavior in the 40-year NCEP–NCAR reanalysis.
J. Climate
,
13
,
873
885
.
Simmonds
,
I.
, and
K.
Keay
,
2000b
:
Variability of Southern Hemisphere extratropical cyclone behavior 1958–97.
J. Climate
,
13
,
550
561
.
Simmonds
,
I.
, and
K.
Keay
,
2002
:
Surface fluxes of momentum and mechanical energy over the North Pacific and North Atlantic Oceans.
Meteor. Atmos. Phys.
,
80
,
1
18
.
Simmonds
,
I.
, and
M. R.
Drinkwater
,
2007
:
A ferocious and extreme Arctic storm in a time of decreasing sea ice. Research Activities in Atmospheric and Oceanic Modelling, Rep. 37, WMO/TD- 1397, J. Côté, Ed., World Meteorological Organization, 2.27–2.28
.
Simmonds
,
I.
,
R. J.
Murray
, and
R. M.
Leighton
,
1999
:
A refinement of cyclone tracking methods with data from FROST.
Aust. Meteor. Mag.
,
June
,
35
49
.
(special edition).
Simmonds
,
I.
,
K.
Keay
, and
E-P.
Lim
,
2003
:
Synoptic activity in the seas around Antarctica.
Mon. Wea. Rev.
,
131
,
272
288
.
Solomon
,
S.
,
D.
Qin
,
M.
Manning
,
Z.
Chen
,
M.
Marquis
,
K. B.
Averyt
,
M.
Tignor
, and
H. L.
Miller
,
2007
:
Climate Change 2007: The Physical Science Basis.
Cambridge University Press, 996 pp
.
Stephenson
,
D. B.
,
V.
Pavan
, and
R.
Bojariu
,
2000
:
Is the North Atlantic Oscillation a random walk?
Int. J. Climatol.
,
20
,
1
18
.
Stephenson
,
D. B.
,
V.
Pavan
,
M.
Collins
,
M. M.
Junge
, and
R.
Quadrelli
,
2006
:
North Atlantic Oscillation response to transient greenhouse gas forcing and the impact on European winter climate: A CMIP2 multi-model assessment.
Climate Dyn.
,
27
,
401
420
.
Stroeve
,
J. C.
,
M. C.
Serreze
,
F.
Fetterer
,
T.
Arbetter
,
W.
Meier
,
J.
Maslanik
, and
K.
Knowles
,
2005
:
Tracking the Arctic’s shrinking ice cover: Another extreme September minimum in 2004.
Geophys. Res. Lett.
,
32
.
L04501, doi:10.1029/2004GL021810
.
Stroeve
,
J. C.
,
M. M.
Holland
,
W.
Meier
,
T.
Scambos
, and
M.
Serreze
,
2007
:
Arctic sea ice decline: Faster than forecast.
Geophys. Res. Lett.
,
34
.
L09501, doi:10.1029/2007GL029703
.
Swift
,
J. H.
,
K.
Aagaard
,
L.
Timokhov
, and
E. G.
Nikiforov
,
2005
:
Long-term variability of Arctic Ocean waters: Evidence from a reanalysis of the EWG data set.
J. Geophys. Res.
,
110
.
C03012, doi:10.1029/2004JC002312
.
Teng
,
H. Y.
,
W. M.
Washington
,
G. A.
Meehl
,
L. E.
Buja
, and
G. W.
Strand
,
2006
:
Twenty-first century Arctic climate change in the CCSM3 IPCC scenario simulations.
Climate Dyn.
,
26
,
601
616
.
Thompson
,
D. W. J.
, and
J. M.
Wallace
,
1998
:
The Arctic Oscillation signature in the wintertime geopotential height and temperature fields.
Geophys. Res. Lett.
,
25
,
1297
1300
.
Thompson
,
D. W. J.
, and
J. M.
Wallace
,
2000
:
Annular modes in the extratropical circulation. Part I: Month-to-month variability.
J. Climate
,
13
,
1000
1016
.
Trigo
,
I. F.
,
2006
:
Climatology and interannual variability of storm-tracks in the Euro-Atlantic sector: A comparison between ERA-40 and NCEP/NCAR reanalyses.
Climate Dyn.
,
26
,
127
143
.
Uppala
,
S. M.
, and
Coauthors
,
2005
:
The ERA-40 Re-Analysis.
Quart. J. Roy. Meteor. Soc.
,
131
,
2961
3012
.
Walczowski
,
W.
, and
J.
Piechura
,
2006
:
New evidence of warming propagating toward the Arctic Ocean.
Geophys. Res. Lett.
,
33
.
L12601, doi:10.1029/2006GL025872
.
Walland
,
D. J.
, and
I.
Simmonds
,
1999
:
Baroclinicity, meridional temperature gradients, and the southern Semiannual Oscillation.
J. Climate
,
12
,
3376
3382
.
Wang
,
J.
,
M.
Ikeda
,
S.
Zhang
, and
R.
Gerdes
,
2005
:
Linking the northern hemisphere sea-ice reduction trend and the quasi-decadal Arctic sea-ice oscillation.
Climate Dyn.
,
24
,
115
130
.
Wang
,
M. Y.
,
J. E.
Overland
,
V.
Kattsov
,
J. E.
Walsh
,
X. D.
Zhang
, and
T.
Pavlova
,
2007
:
Intrinsic versus forced variation in coupled climate model simulations over the Arctic during the twentieth century.
J. Climate
,
20
,
1093
1107
.
Wang
,
X. L.
,
V. R.
Swail
, and
F. W.
Zwiers
,
2006
:
Climatology and changes of extratropical cyclone activity: Comparison of ERA-40 with NCEP–NCAR reanalysis for 1958–2001.
J. Climate
,
19
,
3145
3166
.
Webster
,
P. J.
,
1981
:
Mechanisms determining the atmospheric response to sea surface temperature anomalies.
J. Atmos. Sci.
,
38
,
554
571
.
Wernli
,
H.
, and
C.
Schwierz
,
2006
:
Surface cyclones in the ERA-40 dataset (1958–2001). Part I: Novel identification method and global climatology.
J. Atmos. Sci.
,
63
,
2486
2507
.
Wiese
,
W.
,
1924
:
Polareis und atmosphärische schwankungen (Polar ice and atmospheric fluctuations).
Geograf. Ann.
,
6
,
273
299
.
Winton
,
M.
,
2006
:
Amplified Arctic climate change: What does surface albedo feedback have to do with it?
Geophys. Res. Lett.
,
33
.
L03701, doi:10.1029/2005GL025244
.
Wu
,
A.
,
W. W.
Hsieh
,
G. J.
Boer
, and
F. W.
Zwiers
,
2007
:
Changes in the Arctic Oscillation under increased atmospheric greenhouse gases.
Geophys. Res. Lett.
,
34
.
L12701, doi:10.1029/2007GL029344
.
Wunsch
,
C.
,
1999
:
The interpretation of short climate records, with comments on the North Atlantic and Southern Oscillations.
Bull. Amer. Meteor. Soc.
,
80
,
245
255
.
Yang
,
J.
,
J.
Comiso
,
R.
Krishfield
, and
S.
Honjo
,
2001
:
Synoptic storms and the development of the 1997 warming and freshening event in the Beaufort Sea.
Geophys. Res. Lett.
,
28
,
799
802
.
Yang
,
J.
,
J.
Comiso
,
D.
Walsh
,
R.
Krishfield
, and
S.
Honjo
,
2004
:
Storm-driven mixing and potential impact on the Arctic Ocean.
J. Geophys. Res.
,
109
.
C04008, doi:10.1029/2001JC001248
.
Zhang
,
J.
,
2005
:
Warming of the Arctic ice-ocean system is faster than the global average since the 1960s.
Geophys. Res. Lett.
,
32
.
L19602, doi:10.1029/2005GL024216
.
Zhang
,
X.
,
J. E.
Walsh
,
J.
Zhang
,
U. S.
Bhatt
, and
M.
Ikeda
,
2004
:
Climatology and interannual variability of Arctic cyclone activity: 1948–2002.
J. Climate
,
17
,
2300
2317
.

Footnotes

* Current affiliation: National Meteorological and Oceanographic Centre, Australian Bureau of Meteorology, Melbourne, Victoria, Australia.

Corresponding author address: Ian Simmonds, School of Earth Sciences, University of Melbourne, Melbourne, Victoria 3010, Australia. Email: simmonds@unimelb.edu.au